FINANCING MECHANISMS FOR PUBLIC PRIVATE PARTNERSHIPS: AUSTRALIAN EXPERIENCE Michael Regan1; Jim Smith2 and Peter Love3 12
3
School of Sustainable Development & Architecture, Bond University, Gold Coast Queensland 4229, Australia
School of Built Environment, Curtin University, Perth, WA 6845, Australia.
ABSTRACT Recent events in international capital markets has had major impact on the ongoing rollout of Public Private Partnerships (PPPs) and slowed their use. Capital has been hard to source especially for projects over $300 million, the patronage risk model for economic infrastructure is no longer available, debt is more expensive than it was 18 months ago and credit risk insurance is no longer readily available. This has affected bid markets and slowed the delivery of new infrastructures with longer-term implications for economic performance across the whole economy.
There are two broad approaches that government can take to PPP procurement. Firstly, it may make greater use of alternate procurement mechanisms such as alliance contracting, management outsourcing and traditional procurement. Secondly, it may rethink its role in the PPP process and preserve the model by reducing risk and/or participate in the provision of debt finance. This paper reviews prevailing capital market conditions and state investment evaluation. It examines the empirical evidence to identify the options for state provision of infrastructure and direct financial participation in PPP projects.
Keywords: Public Private Infrastructure, state finance, costs.
INTRODUCTION In Australia, the state has provided most of the continent's economic and social infrastructure from first settlement. It has done this directly with appropriations from budgets, with taxation, by applying user-pays charges, raising public debt or by using a combination of these methods. From Federation, infrastructure was viewed as a public good and with successive federal and state governments, most infrastructure assets and businesses for services in energy, public transport, railways, telecommunications, ports and airports were owned and managed by state agencies or corporations. In the past 10 years, private infrastructure provision has assumed much greater significance and state investment has continued to decline in GDP terms. The difficult capital market conditions of the past 3 years has led to calls for a return to state provision using state debt or tax-preferred bonds and a number of other financing methods. The traditional options for state financing of infrastructure procurement include the budget framework, taxation, user charges and state debt. In the past 40 years, other methods have been attempted - special purpose bonds, state financial assistance (supported debt model in the 1
[email protected] [email protected] 3
[email protected] 2
state of Queensland), and state investment authorities (for instance, the Future Fund and the Queensland Investment Corporation). None of these methods has provided sustainable models for ongoing state financial support for privately-managed infrastructure services. This paper will examine the options for state financing of infrastructure and state participation in public private partnerships with a view to considering how the state can address the difficulties of raising private infrastructure finance at the present time.
THE STATE AS A PROJECT LENDER The state may maintain the PPP procurement model in its present form and assume the role of an arms' length lender to projects. This financing option may be feasible in several situations: 1. When capital markets cannot supply the consortium's project finance requirement (a form of market failure) 2. When borrowers cannot raise the credit insurance necessary to secure credit ratings that minimise the cost of debt capital 3. When the cost of private capital is sufficiently high to adversely affect value for money outcomes. State bonds attract a lower risk premium (spread) than non-government securities of the same maturity and credit rating. At 31st January 2012, the spread between AA rated corporate bonds and Commonwealth bonds stood at 173 basis points (bp), for A rated securities, 253bp and for BBB rated securities, 352bp (RBA, 2013a). Critics of PPPs regularly point to the lower borrowing costs of government compared with the private sector. Indeed, in January 2012, AAA rated corporate bonds with maturities of 1-5 years offered yields in the range of 3.54 3.62% pa. Australian Government 3 year (AAA rated) bond yields were 3.53% and 5 year bond yields were 3.62% pa (RBA 2013b). New South Wales Treasury Corporation bond yields were 4.24% (3 years) and 4.48% (5 years) at 31 January 2012. The spread between state and corporate bond yields is dynamic and moves on a daily basis. However, the average spread for the 11 months to June 2011/12 was 179bp. If the cost of capital was the only element of the value for money outcome, the lower cost of state debt is a decided advantage. However, value for money is both a quantitative and qualitative test. The quantitative test compares competitive private bids against a public sector comparator which is a risk-weighted lifecycle costed model of traditional procurement that takes into account those risks retained by the state and those transferred to bidders. The qualitative test requires critical examination of a proposal with a view to the public interest, sustainability, design amenity, user benefits and improved service delivery (Partnerships Victoria, 2003).
THE STATE AS LENDER AT THE PROJECT LEVEL In the United Kingdom, HM Treasury sought to improve the value for money performance of PPPs by creating a credit guarantee fund (CGF). The fund was created from Treasury capital market borrowings and on-lent to successful PPP consortia with the aim of reducing the cost of capital of the project (Standard and Poor's 2004). The loan takes the form of senior debt guaranteed by consortium bankers and significantly, it is structured in such a way that the incentives attaching to the consortium's lenders, contractors and facility managers remains intact. A variation of the CGF was employed with the South East Queensland Schools project for the Queensland Government. Other jurisdictions are considering this option. However, these models do present several conceptual problems and the CGF approach was shelved in the United Kingdom after two pilot projects (McKenzie 2008; Regan 2008a).
Both state funding methods require the state to select, evaluate and put to market the PPP project and then to provide debt capital to the successful consortium to construct and/or assume a long term investment position with the undertaking. An important feature of both these arrangements is that the state assumes responsibility for both loan and contract administration.
CREDIT GUARANTEE FINANCE (CGF) CGF was introduced in the United Kingdom in 2003 to provide a multi-purpose mechanism for employing public debt capital to PPP projects. In a conventional PPP, the consortium arranges the equity and mezzanine capital requirement (for example, shareholder loans) and raises its debt requirement from the capital market. To lower the cost of debt capital, the SPV will have the project assessed by a credit rating agency (the underlying rating) with a view to obtaining credit enhancement (credit risk insurance) from a monoline agency (the headline rating). For a fee, the SPV will secure an AAA credit rating from the insurer which lowers borrowing costs. The recent sub-prime credit crises has reduced the numbers of active monoline insurers in international and domestic capital markets and many of the major insurers no longer possess the important AAA credit rating (Regan, 2008a: 18). The higher cost of SPV debt over contracts of 20 and 30 years significantly impairs value for money outcomes. The CGF approach has the state providing senior debt to the SPV supported by a monoline agency's AAA guarantee of the consortium's obligations. CGF was financed with an initial £1 billion of debt raised by HM Treasury in 2004 and CGF was trialled with two health projects in 2004 (Leeds) and 2005 (Portsmouth). In the Leeds project, the private consortium's financiers provided the credit guarantee and for the Portsmouth project, the guarantee was furnished by a monoline insurer. An assessment of both projects identified lifecycle interest cost savings to be in the range 8-16% of aggregate finance costs. The nucleus of the transaction is the guarantee furnished by the consortium's bankers or a credit enhancement agency (monoline insurer) to the state as security for the loan. The objective of CGF is to reduce the consortium's cost of capital and thereby improve the long-run and overall value for money outcomes for the state. This arrangement is a departure from traditional project finance principles whereby senior debt is limited recourse and secured over underlying project assets. CGF is, in fact, full recourse debt and this does affect the traditional incentive mechanisms that are a feature of conventional project financings (HM Treasury 2005). The CGF modifies the underlying PPP transaction in a number of important ways. PPP transactions rely on a combination of incentives, regulatory and governance frameworks and, market discipline. Against this background, the effects of CGF are as follows:
CGF gives the state three roles in a PPP transaction - project origination, project finance and contract administration during the term of the contract. The multiple roles create contractual complexity and potential conflict of interest. For example, a discretionary or disputed abatement of a unitary (or asset availability) payment by a contract administrator may impair the SPVs capacity to service debt leading to default and a call under the guarantee (NAO 2003).
The substitution of state for corporate debt effectively reduces the SPVs cost of capital. However, the benefit is reduced by direct and deadweight costs, and transaction costs. Additionally, for the state to assume transactional and guarantee risk, the interest rate
should also recognise a credit risk premium notwithstanding an unconditional guarantee. Corporate AAA debt attracts a higher spread than state debt of that standing and a corporation cannot have a credit rating superior to that of the sovereign government.
Application of CGF requires Treasury to assume the role of an arm's length lending bank involving loan documentation and administration, legal and advisory fees, regulatory oversight and industry-specific technical knowledge and the agency costs involved. These transactional costs are duplicated by the private providers of guarantees.
CGF introduces another layer of contractual complexity into the PPP transaction which contributes to additional decision-making friction that may incur time and cost delays.
In the volatile market conditions of the past 18 months, a number of leading monoline insurers received credit rating downgrades and several have withdrawn from the market altogether (Regan 2008b). In August 2008, only 18% of the domestic credit insurance market had held their Standard and Poor's AAA credit ratings of 2006. This suggests that the void will need to be filled by consortium members and/or domestic banks with associated guarantee fees and transaction costs that would offset the cost of debt savings. The underlying credit rating of PPP projects in Australia is generally BBB or investment grade reflecting project fundamentals and the credit strength of the principal contractors as a proxy for the likelihood of satisfactory project delivery. This raises the question of a shadow credit risk premium suggesting the impact of the CGF in prevailing market conditions may well be negligible.
PPP consortia are generally a collection of entities with different incentives and timing objectives. Few corporations plan much beyond 5 years and most for much less than that. Construction companies have an appetite for delivery risk but are averse to long-term equity holding positions. Portfolio investors prefer stable, long-term revenue streams which favour projects with mature income and expense characteristics and a low risk profile. SPV members may migrate across different industry sectors attracted by diversification, higher returns or simply greener pastures. The dynamics of the listed market favour flexibility. The CGF model with its long term debt commitments inhibits this flexibility, which may reduce depth in bid markets.
PPPs are an incomplete contract - commercial and financial settings change, risk profiles are dynamic, opportunity may arise for renegotiation of parts of the agreement, real and embedded options may be exercised, and there will be ongoing changes to the marginal return on investment and underlying investment economics. Long-term debt arrangements may inhibit sponsor flexibility.
Economies of scale suggest that for the CGT program to derive large scale benefits for the state and mitigate unsystematic risk, it would need to be applied to a large number of industry-specific projects. A further criticism of the CGF model is that it doesn't offer the incentive mechanism available with conventional PPPs whereby senior debt providers possess step-in rights in the event of default and possess incentives to negotiate a commercial and operational rescue of the project whilst maintaining service delivery. Under CGF, the incentives are less clear.
The CGF model provides a substitute for private debt and therefore maintains the important incentive framework for the members of the SPV to perform under the contract. However, it
does away with the independent private financier and removes an important performance monitoring and governance check on the SPV. A bank lender advancing senior debt to a SPV holds a limited recourse security that gives step-in rights in the event of default. The lender assumes an independent financial monitoring role with a view to minimising non-compliance with KPIs that result in abatements, penalties and reduction in debt servicing capacity. Additionally, private lenders bring market disciplines, know-how, financial risk-management expertise and industry experience to the role and apply a further level of governance. The CGF replaces the bank with financial institutions that don't bring the same commercial acumen or experience to the role. A passive debt guarantor and a Treasury Corporation will be required to deal with performance monitoring, loan administration, a dynamic risk environment, financial risk management and the financial economics of a long-term PPP. These activities are not always central or core competencies. The CGF model is better suited to the European SPV structure than the variation employed in Australia. European SPVs are generally contractor-led and employ long-term project finance arrangements with embedded refinance options. In Australia, SPVs are often led by financial intermediaries who provide the capital underwritings for bids. CGF was not used after the Leeds and Portsmouth hospital PFI contracts although comprehensive guidance and standard form documentation has been put in place (HM Treasury2003). There is no commitment to proceed further with CGF although it remains an option for the future.
THE SUPPORTED DEBT MODEL In 2008-09, the Queensland Government conducted a pilot program for a PPP in the education sector using a hybrid variation of CGF described as the supported debt model (SDM). The SDM has several distinguishing characteristics: 1. The SPV arranges private construction finance 2. When the asset is commissioned, the state provides a long-term finance facility to repay construction finance 3. The level of state debt employed is calculated using a formula that equates to a minimum asset value (or recoverable amount) in the event of consortium default. This may be expressed as a percentage of on-completion value. The state assumes the role of limitedrecourse lender although the arrangement does not rule out a requirement for full and partial guarantees. 4. The state holds the senior debt position. The SPV will raise additional subordinated debt and equity capital from private sources. The SDM preserves traditional ex ante incentives and does not require credit enhancement or supporting private guarantees. 5. The lower cost of state debt reduces the cost of capital for the SPV and improves value for money lifecycle finance costs which should be reflected in an improved value for money outcome for the state. The SDM takes advantage of the significant change in risk profile that accompanies the commissioning of a PPP project. The SDM is calculated against a notional risk-free minimum value for the project against which the state can make debt finance available to the project at cost. The SDM has three distinctive characteristics:
SDM financing is attractive from a value for money perspective, particularly given the recent increased spreads for private debt following the global financial crisis and it avoids the costly requirement for credit insurance.
The SDM model contributes to high transactional costs during the early stages of the project although these may decline when the project reaches operational maturity. Overall contractual friction should be less for SDM than CGF with lower transaction and agency costs.
SDM has parallels with conventional project finance but shares little in common with the short to medium-term corporate finance employed in most Australian PPPs. An implication of the model that may adversely affect improved value for money outcomes is the need for privately sourced junior and mezzanine debt or equity capital to bridge the gap between the recoverable amount and the higher level of senior and subordinated debt usually sourced from banks. Subordinated debt carries higher risk spreads. Recent research (McKenzie, 2008) suggests that the average state contribution to PPP debt capitalisation is around 65-70% of capital requirement indicating a mezzanine/junior debt participation of around 15-20% in addition to the equity contribution. The overall cost of debt will be determined on a project basis and particularly on the underlying credit strength of the underlying transaction and the quality and experience of the consortium. The use of higher levels of private mezzanine/subordinated debt and equity capital in prevailing market conditions will increase the SPV's cost of capital and offset part of the SDM's savings in lower debt cost without the relief offered by revaluation and refinancing. However, when capital is difficult to source, this is less of a consideration provided the VFM outcome remains positive. The break-even point for SDM is narrow and estimates suggest that this may occur when average private debt spreads exceed 500 basis points (McKenzie, 2008). Depending on the unsystematic risk profile of the underlying transaction, this is most likely to occur in prevailing market conditions. SDM may raise the sponsor's overall cost of capital and this could offset a significant part of the cost savings achieved with lower cost senior state debt. A second issue is the consortium's lack of flexibility. The SDM removes the short and medium term revaluation and refinancing gains of the Australian approach to long-term PPP contracts. PPPs are long-term incomplete contracts frequently containing embedded options to deal with changed operational or broader network conditions. Revaluation enables early-stage risk-taking equity investors to exit the project and sell down to more risk-averse fund managers and institutional investors (Tapper and Regan, 2007). Refinancing has several important advantages for mature projects - it permits an increase in senior debt (thereby reducing more costly subordinated debt and overall cost of capital), and facilitates higher leverage and a withdrawal/return to equity. Refinancing gains are shared with the state under Australian PPP guidelines. A third issue is the additional administrative cost that SDM imposes on the state. As a secured lender to the project, the lending agency will need to replicate private banking credit assessment and loan administration roles. This will add significantly to transaction and agency costs and in the nature of government, adds a layer of procedural and governance friction. The SDM was used for the first time in the South East Queensland Regional School projects in 2010.
SUMMARY The SDM and CGF meet two important needs in present market conditions. First, both approaches preserve the PPP procurement method and permit its continued evolutionary development. Second, they offer a form of state project participation and risk sharing in projects. Both models preserve the basic characteristics and advantages of PPPs - risk transfer, an output specification, innovation and private incentives. Nevertheless, both approaches raise several potential difficulties. The first concerns the sustainability of lower cost of capital after adjusting for higher levels of equity or mezzanine capital and the friction costs associated with loan administration. The second concerns direct and indirect costs. The state will finance these projects by borrowing in capital markets, raising taxes or sourcing the capital from existing appropriations. The first two methods attract deadweight costs and the third carries an opportunity cost. The second concerns the removal of capital markets’ discipline normally associated with bank or bond finance. PPPs rely on symmetrical incentive frameworks - bankers seek to minimise the probability of default and put in place comprehensive reporting, governance and monitoring systems to administer the loan. These include performance criteria such as debt to value profiles, contributions to sinking funds and reserve accounts, debt service coverage ratios and cash flow management. This process operates quite independently of the operational performance indicators agreed with the state under the franchise agreement. The state monitors performance under its contract management framework and this continues for the term of the loan. This is essentially the monitoring of operational performance although a comprehensive contract management framework will include matters at a corporate level that may affect a company's participation in a long-term contract. This is essentially a relationship management role that that is framed around monitoring of operational performance and the state of health of the SPV and its members (Partnerships Victoria, 2001). Each of the central parties to the PPP arrangement, the state, the consortium or SPV and lenders, are contractually linked in a tripartite agreement. If the SPV defaults under either the PPP or the loan agreements, the lender holds step-in rights to secure the asset and maintain service delivery. Failure to preserve this incentive framework will affect the long-term performance of PPPs. The removal of capital markets disciplines affects the incentive framework of the PPP and reduces investor flexibility. This is a greater concern in economic infrastructure projects where investors are exposed to full or partial patronage risk. In social infrastructure projects, the scope for revaluation gains is reduced if revenue takes the form of a capital charge or availability payment. Nevertheless, the existence of abatement and incentive payments in the early years’ operation may lead to high investor return volatility and refinancing gain (NAO, 2005). This may have an adverse effect on private sector appetite for PPP projects in the future and depth in competitive bid markets. Finally, the CGF and SDM require the state, as a project lender, to administer the loan, ensure adequate management practices are in place and regulate the PPP arrangement. This activity is not a core skill of state treasury corporations. Nor is it a task for which they possess the personnel, the experience and the know-how. This may add significantly to transactional friction and cost which should be explicitly recognised in the value for money evaluation.
REFERENCES H.M. Treasury (2003) The Green Book, Appraisal and Evaluation in Central Government, Treasury Guidance, The Stationery Office, London.
McKenzie, R.. (2008) Strategies to Improve Value for Money in Financing Public Private Partnerships, Public Infrastructure Bulletin, Edition 7, pp. 5-16. National Audit Office (2003) The Operational Performance of PFI Prisons, Report by the Comptroller and Auditor General, HC 700 Session 2002-03, 18 June. National Audit Office (2005) Improving Public Services Through Better Construction, Vols 1 & 2, Report by the Comptroller and Auditor General, HC209, Session 2004-5, 10 February 2005. Partnerships Victoria (2001), Practitioners Guide, Guidance Material, Department of Treasury and Finance, Melbourne. Partnerships Victoria (2003), Public Sector Comparator, Technical Note, Department of Treasury and Finance, Melbourne. Regan, M. (2008a) What Impact Will Current Capital Market Conditions Have On Public Private Partnerships? A report for the Infrastructure Association of Queensland, Research Report 121, School of Sustainable Development, Bond University, Gold Coast, Queensland. Regan, M. (2008b) Public Private Partnerships, What Lessons Have We Leamt? Working Paper WP100, School of Sustainable Development, Bond University, Gold Coast, Queensland. Reserve Bank of Australia, (2013a) Staistics, Table F3, Sydney. Reserve Bank of Australia (2013b) Statistics, Table F2, Sydney. Standard and Poor's (2004) Public Finance/Infrastructure Finance, Credit Survey of the Private Finance Initiative and Public Private Partnerships, New York, May. Tapper, P. and Regan, M. (2007) Australia’s Hybrid Approach to Project Finance, Public Infrastructure Bulletin, Edition 6, pp. 21-24.
